The concept of the text--as the product of a sign system that must be interrogated--has been extremely productive for literary and cultural studies, and Barthes was much involved in the promotion of the concept. The best-known essay about the concept of the text from this period is no doubt his "De l'oeuvre au texte." Although the ideas it draws on were certainly in the air--none of the characterizations of the text were at all surprising--Barthes's distinctive articulation of them is highly idiosyncratic and not conducive, as Barthes himself might be the first to admit, to the advancement of methodological clarity or of an analytical program in literary and cultural studies. Barthes locates the rise of the concept of texte in the context of interdisciplinary encounters, as both a product of this situation and a response to interdisciplinarity, and he insists that we are not dealing with a radical mutation, but with a "glissement epistémologique plus qu'une véritable coupure" (an epistemological slippage more than a veritable break).

Barthes describes the notion of the text through an opposition between text and work on a number of different parameters. The persistence of this opposition gives his essay its clarity and force: the text is always being opposed to the work. But what Barthes seems to wish to resist above all is the idea that the concept of text can be one that replaces the concept of work; for him, it isn't a matter of changing our view of objects previously treated as works and conceiving them instead in a new way, tempting though this idea might be. Rather, it is as though the text is a new thing, previously unglimpsed. What constitutes the text, for instance, is "sa force de subversion à l'égard des classements anciens" (its force of subversion with regard to the old classifications). Text is "ce qui se porte à la limite des règles de l'énonciation, (la rationalité, la lisibilité, etc.)" (what is situated at the limits of the rules of speech act [rationality, readability, etc.]). It is irreducibly plural, a practice or play of the signifier generating the infinite deferral of the signified; it is not consumed by the reader but solicits collaboration on the part of the reader and functions not as an object of consumption yielding plaisir but as a practice of disruptive and self-disruptive jouissance.

Barthes distinguishes, as his title tells us, oeuvre from texte. But his claim is not that literary studies used to operate with one notion of its object, that of the work, which now it has reason to contest, so that it now thinks of the objects of its study as texts, with myriad consequences. On the contrary, he wants to insist that there are indeed oeuvres, things that we continue rightly to describe and analyze as such, which are locatable, describable; and then there is texte. Although he speaks here and there in the essay of texts in the plural, he insists that texts are not countable, computable, or locatable; he generally speaks of le Texte (singular, with a capital T), and maintains that there is "du Texte" (some text) to be located here and there in oeuvres. The texte en soi, pure text, is not something that can be found, analyzed; so while we can say that some writings are oeuvres, we can't say that other writings are texts--only perhaps, that they have some text about them. And Barthes insists from the outset, knowing that this is how we are inclined to interpret his essay, "Il ne faut pas se laisser aller à dire: l'oeuvre est classique, le texte est d'avant garde" ("we must not permit ourselves to say: the work is classical, the text is avant-garde"). We mustn't because there can be "du Texte dans une oeuvre très ancienne, et bien des produits de la littérature contemporaine ne sont en rien des textes" ("there can be some Text in a very old work, and many products of contemporary literature are not texts at all"). But of course this argument reinforces the idea that the oeuvre is something like "normal literature" and text is something avant-garde--just so radical that it can't be pinned down.

In an essay of 1972 entitled "Jeunes chercheurs" ("Young Researchers") Barthes writes,

[W]hen we say the Text it is not in order to make it divine, to make it the deity of a new mystique, but to denote a mass, a field requiring a partitive and not a numerative expression: all that can be said of a work is that there is Text in it.

But this usage certainly does make Texte an honorific concept, if not quite a God. And indeed, my principal objection to Barthes's formulations in "De l'oeuvre au texte" is that while remaining within a logic of opposition, they work to generate a mystique of the text: it is something so radical, disruptive, indeterminate, that it is not even an object but a practice or process, at best identifiable in certain moments. Barthes describes accurately, I think, what was actually happening: jeunes chercheurs, as he notes, see it as their task to "repérer ce qu'il y peut avoir de Texte dans Diderot, dans Chateaubriand" (to explore what Text there can be in Diderot, in Chateaubriand). It is a matter of finding "ce qui, dans l'oeuvre ancienne, est Littérature et ce qui est Texte" (what in old works is Literature and what is Text). If you can show that there is "du Texte" in works of the past, you have shown that they are radical, exciting, worthy of attention.

Jonathan Culler: "Barthes, Theorist," The Yale Journal of Criticism 14 (2001): 439-446; quote from pp. 441-44.